Model-based analysis of medical images
Sophisticated medical imaging methods have become ever more widely available, resulting in a flood of data and creating a demand for medical image analysis methods to extract information automatically.
Before our research, image segmentation applications were created in an ad hoc way, using application-specific algorithms that were labour-intensive to develop and often unreliable. We introduced a completely new paradigm, demonstrating that anatomical knowledge could be captured systematically from a training set of images, and used in an entirely generic way to locate structures of interest in new images, with high accuracy and reliability. The methods we developed have been very widely adopted, leading to both economic and healthcare impacts via both specialist small/medium size enterprises (SMEs) and multi-national companies in medical imaging, pharmaceuticals, and orthopaedics.